Is it just me? Can anyone confirm success using Olaf Kirch's latest nfs
package (kernel-based server & async write support) under later kernels?
It appears to work fine for small transactions, but all transfers
involving > 1-Meg (or thereabouts) result in a dead server - and
occasionally a dead box.
I keep reverting to 2.1.36, which exhibits no such problems.
This is on an otherwise-stable box with no history of hardware flakiness.
Steve